If you’re an affiliate marketer, you probably understand how the basic model works: You sign up to be an affiliate of, say, Amazon and earn monthly commissions based on the number of visitors you send to Amazon who complete a purchase.
The advertiser pays you an amount generally based on a percentage of the purchased item’s price.
Advertisers generally pay their affiliates following the “last click” system, meaning they pay the marketer most directly tied to the sale. In other words, they pay the one whose website brought a consumer directly to Amazon where the purchase was then completed.
Adding complexity to the picture is the fact that larger advertisers tend to use multiple affiliate marketers—in addition to running “paid search” campaigns, as well as social media advertising in the form of a Facebook ad, for example.
Now consider this: The consumer, en route to purchasing the product, could easily click on a number of different affiliate marketers’ links before finally making an actual purchase.
Is it fair then that you earn nothing off a transaction if you were, say, the first click in a chain of clicks that lead to the ultimate purchase?
This is where so-called attribution modeling comes into play. It’s a method of viewing historical data that could very well earn you a percentage of the commission on a sale even if you were not the last click.
“The consumer may touch on multiple marketing channels in their journey to making a buying decision and ultimately buying,” Todd Crawford, co-founder and VP of Strategic Initiatives at Impact Radius, told Small Business Trends in an exclusive interview.
You, as an affiliate marketer, have much to gain from the use of attribution modeling because it is a system of looking at data that values you for whatever particular role you play in an online sales transaction that included your affiliate link, whether first or last.
“Attribution gives the advertiser insights into the contributions that affiliates are generating,” Crawford says. “For affiliates that rarely win [in a last click model] but are involved higher up in the funnel, an advertiser can decide to increase their commission rate so it compensates them for the conversions they do get credit on. The advertiser could also give them a spot bonus to recognize their contributions.”
The problem is that affiliates generally don’t know if they are losing out to other channels.
“They only know how much traffic they sent the advertiser and the conversions [for which] they were credited. If a conversion isn’t credited to them, they have no insights into that,” Crawford explains.
So what can an affiliate marketer do?
“It is probably prudent to inquire as to how the advertiser is tracking and crediting conversions,” Crawford says. “Is it based on last click, can other marketing channels trump an affiliate click, etc?
“They may also ask if the advertiser is doing attribution modeling and how does it change their perspective on the affiliate channel and the specific affiliate partnership.”
Crawford, described as legendary for his role in driving awareness of attribution modeling, discussed the concept in affiliate marketing at the annual AM Days conference in San Francisco in April.
He adds that with attribution modeling, “the sale already happened. Marketers are trying to understand, what was my total cost? Attribution is modeling. You have all this data and you want to learn more.”
Attribution analyzes all the touchpoints with which a consumer came into contact before purchasing an item — and then values each touchpoint accordingly.
For example, say an advertiser utilizes an affiliate program, paid search and a Facebook ad in a multi-channel marketing effort. Under attribution modeling, the advertiser will know, based on historical data, whether his allotment for each of the three channels is optimized based on the advertiser’s goals, which may range from generating more revenue to generating more traffic.
“It’s just another way of looking at the data,” Crawford says.
“I might say, instead of paying 5 percent to this affiliate, I should pay them 4 percent. Or maybe I should pay them 8 percent. Attribution helps me understand all my other marketing costs” so I’d know if a component of my strategy is too expensive or cost efficient, for example.
Most advertisers use “last click” crediting logic, Crawford notes. If they then applied attribution modeling, they’d know, based on their given criteria, “what is working and what is not.”
“It’s an analysis. Should I spend more or less on Facebook, or with an affiliate?”
Using attribution modeling, the advertiser will know, based on its historical data, whether its allotment for each of its marketing channels (use of affiliates, paid search, social media advertising) is optimized, based on the advertiser’s goals, which may range from generating more revenue to generating more traffic.
Businesses are aware of attribution, Crawford says. It boils down to the question: When are they ready to start using it.
They certainly won’t be ready unless they spend a significant amount of revenue on online marketing efforts. Otherwise, “there is not enough complexity for attribution data to deliver much value,” Crawford notes.
A company spending $100,000 per year on online marketing efforts could see a benefit from attribution modeling, but it wouldn’t be that significant. They could realize a savings or increased revenue of around 3 percent, Crawford says.
But if that figure were $100 million, “that is significant,” Crawford notes. Then spending $10,000 to $50,000 a month on attribution solutions would be well worth it—as the company could save or earn additional millions of dollars.
Companies spending millions on online marketing efforts may eventually cultivate a need to understand where the opportunities are. These are the likely candidates that will turn to attribution modeling, to help take their online marketing efforts to the next level.
Once a company decided to make the leap, they can deploy attribution modeling in a variety of ways. It could create an attribution modeling method based on it owns internal data.
Or it could hire a consulting firm that offers the service.
There are different attribution models, as well, Crawford adds. The “decay model,” for example, goes from the last to first touchpoint, “giving less and less credit” as we go from “last click” to first.
Or, a company could decide to credit all touchpoints equally.
“You can look at different models and try to make sense of which one makes best sense for you,” Crawford suggests. It is important to note that attribution modeling is about analyzing data after the fact—not determining or adjusting affiliate commissions dynamically.
Earning some money based on attribution modeling is certainly better than earning no money as per the “last click” mindset that now dominates online marketing.